SPSS chart question

[Jeppe Christensen]

Hi all

This fall under the category stupid questions, even though I'm told
there are none.

I'm making my bar charts in excel at the moment because SPSS seems
unable to do what I want, and that just ain't right:)

I have multiple variables with measurements for different chemicals,
they are all measured in the same unit, and their levels are similar,
so I want to create a single graph (bar chart or boxplot) with their
medians. My question then, is this:

Is it possible to have more than one variable on the X-axis (one for
each chemical)?

Regards
Jeppe

PS: I RTFM:)

[Andy W]

This seems pretty simple in the legacy type graphs. Below is an
example using the dataset I have open at the moment. "FKeyNum" and
"RspdntID" are the numeric variables I am summarizing.

***************************************************************.
EXAMINE VARIABLES=FKeyNum RspdntID
/COMPARE VARIABLE
/PLOT=BOXPLOT
/STATISTICS=NONE
/NOTOTAL
/MISSING=LISTWISE.

GRAPH
/BAR(SIMPLE)=MED(FKeyNum) MED(RspdntID)
/MISSING=LISTWISE.
***************************************************************.

If this isn't sufficient you could make more complicated charts using
GPL, although it would be best for you to describe what your goals are
more explicitly and give an example format your data is currently in.

Obligatory note to avoid "Dynamite Plots" if you plan on adding error
bars to the median estimates for the bar plot!
http://biostat.mc.vanderbilt.edu/wiki/pub/Main/TatsukiKoyama/Poster3.pdf

Andy